Jichu yixue yu linchuang (Jun 2023)

Prediction of five types of general surgical complications based on Logistic regression model

  • CHEN Wangyue, XUE Fang, HAN Wei, WANG Zixing, JIANG Jingmei, YU Xiaochu

DOI
https://doi.org/10.16352/j.issn.1001-6325.2023.06.0974
Journal volume & issue
Vol. 43, no. 6
pp. 974 – 980

Abstract

Read online

Objective To describe the occurrence of general surgical complications and establish prediction models for different types of complications based on a multicenter cohort study. Methods Based on Modern Surgery and Anesthesia Safety Management System Construction and Promotion(MSCP), patients who underwent general surgery in 4 hospitals from January to June 2015 and from January to June 2016 were selected as participants, and perioperative data of patients were collected. Logistic regression was used to identify risk factors and predict complications. Results Among 19 223 patients, 830(4.32%) had complications. Among participants who had complications, 371(44.70%) had incision complications, 190(22.89%) had fistula complications, 310(37.35%) had infection complications, 161(19.40%) had failure complications, and 104(12.5%) died. There were significant differences in the risk factors of different types of complications. The risk factors of incision,infection and failure covered the whole perioperative period, while fistula complications mainly focused on the difficulty of surgery and postoperative treatment, and for death outcomes, postoperative risk factors were more severe than preoperative risk factors. The areas under the curves of prediction were between 0.80-0.93. Conclusions In general surgery, different types of complications have different risk factors. The targeted prediction model can avoid the rough simplification of complications and fuzzy influence of each factor, and can provide reference for the prevention of complications.

Keywords